Dear Yale Daily News,
Mr. Chua’s article spotlights the dangers and limitations of data. We have for decades used the phrase “garbage in, garbage out” as a shorthand expression of this concern. However, with the contemporary metamorphosis of artificial intelligence from imperfect tool to cult status, the tendency is to forget that artificial intelligence is not remotely similar to human intelligence. Indeed the term AI is a deceitful misnomer that equates digital manipulation with cognitive consideration. AI is no more “artificial intelligence” than a rubber garden hose is an “artificial human artery.” True, both are tubes and carry liquids, but their functions and efficiencies are completely different.
An excellent discussion of this point can be found in a Feb. 3 Scientific American article about automated vehicles . In response to numerous “studies” commissioned by various self-driving car manufacturers showing that autonomous cars are safer than human-driven ones, Scientific American points out in detail that the data of these studies is misleading, incomplete, and … well … artificially rather than actually intelligent.
As Mr. Chua notes, data veneration abhors the individual, the different, the rebellious. So do governments. Governments find us pesky individuals harder to control than uniform groups. Thus, policy driven by data tends toward tyranny, whereas policy driven by moral imperatives tends toward empathy. The more AI is seen as a technological savior rather than as a mislabeled tool, the faster we will slide down the slippery slope of ignorance toward philosophical entropy.
James Luce ’66